J. Phys. Chem. B 2007, 111, 9281-9287
9281
Behavior of Electrogenerated Bases in Room-Temperature Ionic Liquids Sarah O’Toole, Sreekanth Pentlavalli, and Andrew P. Doherty* School of Chemistry and Chemical Engineering, DaVid Keir Building, Queen’s UniVersity of Belfast, Stranmillis Road, Belfast, Northern Ireland, BT9 5AG, U.K. ReceiVed: March 27, 2007; In Final Form: May 21, 2007
The reductive electrochemistry of substituted benzophenones in the aprotic room-temperature ionic liquid (RTIL) 1-butyl-1-methylpyrrolidinium bistriflimide occurs via two consecutive one-electron processes leading to the radical anion and dianion, respectively. The radical anion exhibited electrochemical reversibility at all time-scales whereas the dianion exhibited reversibility at potential sweep rates of g10 V s-1, collectively indicating the absence of strong ion-paring with the RTIL cation. In contrast, reduction in 1-butyl-3methylimidazolium bistriflimide is complicated by proton-transfer from the [Bmim] cation. At low potential sweep rates, reduction involves a single two-electron process characteristic of either an electrochemical, chemical, electrochemical (ECE) or disproportion-type (DISP1) mechanism. The rate of radical anion protonation in [Bmim] is governed by basicity and conforms to the Hammett free-energy relation. At higher potential sweep rates in [Bmim][NTf2], reduction occurs via two consecutive one-electron processes, giving rise to the partially reversible generation of the radical anion and the irreversible generation of the dianion, respectively. Also, the redox potentials for the reversible parent/radical anion couples were found to be a linear function of Hammett substituent constants in both RTIL media and exhibited effectively equivalent solvent-dependent reaction constants, which are similar to those for reduction in polar molecular solvents such as acetonitrile or alcohols.
Introduction Over recent years, room temperature ionic liquids (RTILs) have become important solvents for organic chemical reactions,1,2 support media for homogeneous3,4 and heterogeneous catalysis,5,6 and electrochemistry.7-10 The key advantage attributed to RTILs is their lack of volatility and hence their ease of containment and low environmental/hazard impact, relative to molecular solvents.1,2 Also, it has become clear that RTILs, depending on composition, can influence chemical reactivity including the acceleration of reaction rates11 and controlling the outcome of reactions.12 As such, they are viewed as attractive alternatives to conventional organic solvent systems such as acetonitrile, tetrahydrofuran, dimethylsufoxide, etc. for industrial applications.1,2 Because RTILs are inherently ionic conducting and generally electrochemically robust, they are also emerging as important “solvents” for direct7-9 and mediated organic electrochemical transformations.10,13 In particular, their inert nature renders RTILs useful for studying radical chemistry; for example, it has been shown that the electrochemically generated and highly reactive superoxide species (O2•-) is stable in ionic liquids.14-16 Investigations of preparative and mechanistic organic electrochemistry in RTILs include the Ni-catalyzed homocoupling of organic halides;10 the oxidation of anthracene, naphthalene, durene, 1,4-dithiafulvene, and veratrole;17 the reduction of benzaldehyde,9,18 dinitrobenzenes,19 9-chloroanthracene, 4-chlorobenzophenone,20 acetophenone,21 and nitrocompounds.22 The transfer of organic electrochemical processes, frequently involving reactive radical species, from conventional electrolytic media to RTILs may not necessarily be straightforward, since * To whom correspondence should be addressed. E-mail: a.p.doherty@ qub.ac.uk. Phone: +0044 (0) 2890 974481. Fax: +044 (0)2890 976524.
it is believed that the electrogenerated radical anions may interact with the ionic liquid supporting medium altering chemical reactivity.19,22 For example, the reductive electrochemistry of dinitrobenzenes in [Bmim][BF4] ([Bmim] ) 1-butyl-3-methylimidazolium, [BF4] ) tetrafluoroborate) occurs via a single two-electron process whereas, in acetonitrile, reduction occurs via two consecutive one-electron processes,19 behavior which was attributed to strong ion pairing between the electrogenerated dianion species and the [Bmim] cation. Also, the reductive dehalogenation20 of 9-chloroanthracene and 4-chlorobenzophenone in various RTILs was “considerably affected” by the RTILs medium, where the effect is dependent on the charge delocalization of the radical anion. It was reported that, when the radical anion charge is significantly delocalized, bond cleavage rates were greatly accelerated in RTILs, whereas when charge is less delocalized, cleavage rates were a factor of 10 lower in RTIL media, relative to acetonitrile. In contrast, the follow-up chemistry of electrogenerated cation radicals of anthracene, naphthalene, and durene and the electrodimerization/electropolymerization of 1,4-dithiafulvene and veratrole17 in imidazolium and quaternary ammonium-based ionic liquids were found to be mechanistically the same as in acetonitrile. Although follow-up chemistry may, or may not, be the same in RTILs, relative to aprotic polar solvents, two common effects of RTILs are evident. First, mass transport rates (reported as diffusion coefficients) are smaller by a factor of between 10 and 100 in all RTIL media, which is due to the viscous nature of the media. Also, it appears that heterogeneous electrontransfer rate constants (k0) (in imidazolium-based ionic liquids at least) can be smaller by as much as 2 orders of magnitude for processes controlled by outer-sphere dynamics.22 This effect was attributed to the “special” solvent effect of RTILs.
10.1021/jp072394n CCC: $37.00 © 2007 American Chemical Society Published on Web 07/14/2007
9282 J. Phys. Chem. B, Vol. 111, No. 31, 2007
O’Toole et al.
Figure 1. Structures on (a) [Bmim], (b) [Bmpyrd], and (c) [NTf2] ions.
In this contribution, we examine the detailed electrochemistry of a range of substituted benzophenones in two different RTIL media, which are 1-butyl-3-methylimidazolium bistriflimide ([Bmim][NTf2]) and 1-butyl-1methylpyrrolidinium bistriflimide [Bmpyrd][NTf2]. The RTILs were chosen on the basis that [Bmim] is protic23 whereas [Bmpyrd] is not, in the knowledge that benzophenone electrochemistry is sensitive to proton availability.24 Also, the reduction potentials of the substituted benzophenones depend on the electronic nature of the substituents and are sensitive to the polarity of the electrolytic media;24 therefore, information concerning the polarity of these media should be evident through the Hammett free-energy relation.25
Figure 2. Cyclic voltammograms of the first reduction of BP4 in [Bmpyrd][NTf2] at sweep rates of (a) 0.010, (b) 0.025, (c) 0.050, (d) 0.075, and (e) 0.100 V s-1.
the effects of electrode fouling can be observed after several voltammetric cycles; it is therefore imperative to perform measurements with freshly polished electrodes. The reference electrode was the Ag/AgCl/sat. KCl electrode fitted with a 1 cm long ceramic frit to prevent leakage of H2O into the RTIL medium. The redox potential for the ferrocene/ferricinium redox couple, in both [Bmpyrd][NTf2] and [Bmim][NTf2], is 0.320 ( 0.010 V relative to this reference electrode. Prolonged immersion of this reference electrode did not alter the voltammetry recorded, relative to either the initial voltammetry or to voltammetry performed with a solid-state reference electrode. Cyclic voltammetry and chronoamperometry were performed at 293 ( 1.0 K under a dry N2 atmosphere. Positive-feed-back iR compensation was used throughout. Typically, compensated resistances were 0.5-1.0 kΩ, depending on electrode positioning.
Experimental Section [Bmpyrd][NTf2] and [Bmim][NTf2] RTILs (chemical structures are shown in Figure 1) were prepared by ion metathasis of the corresponding chloride and lithium salts in acetonitrile (ACN) as previously described.26 The crude ionic liquids were washed 10 times with deionized H2O to remove residual LiCl, whereupon the washed ILs were dissolved in ACN containing activated carbon to remove residual organic impurities. The resultant liquids were colorless. Prior to use, the liquids were dried on a Schlenk line at 80 °C under high vacuum for at least 24 h. Typically, this procedure ensures a residual water concentration of e10 parts per million (ppm). So-dried ionic liquids were transferred anaerobically to a dried, airtight electrochemical cell under a N2 blanket The benzophenones examined are listed in Table 1 along with their Hammett substituent constants (∑σx). Note that the benzopheneones listed 1-4 will be referred to as BP1, BP2, BP3, and BP4 in the rest of the manuscript. Because of the viscous nature of RTILs, it is impossible to dispense accurate volumes of liquid; therefore, substrates solutions were prepared gravimetrically, with typically 15 g of dried RTIL being used in the electrochemical experiments. Electrochemical measurements were performed in the threeelectrode configuration using a BAS CW-50 potentiostat. The working electrodes were glassy carbon (3 mm diameter (BAS)) which were polished with 0.05 µm alumina as an aqueous slurry prior to use. It is our experience with BP electrochemistry that
Results and Discussion Electrochemistry of BP1-4 in [Bmpyrd][NTf2]. Figure 2 shows cyclic voltammograms (CVs) for the reduction BP4 at glassy carbon (GC) electrodes in [Bmpyrd][NTf2], at potential sweep rates (ν) from 0.010 to 0.100 V s-1. The same electrochemistry was also observed for BP1-3 in this medium. This reduction can be attributed to the well-known24,27 oneelectron process generating the radical anion species; the halfwave potentials (E1/2,1,Bmpyrd) for BP1-4 are given in Table 1. The electrochemical reversibility of the reduction process is evident by the clear presence of a reoxidation process at very low potential sweep rates. However, at the lowest sweep rate a slightly smaller (approximately 15-20%) anodic reoxidation current is evident. This behavior is probably due to a very slow
TABLE 1: Benzophenones Examined along with Abbreviation, Hammett Substituent Constants (∑σx), Half-Wave Potentials for the First Reduction (E1/2,1), and Diffusion Coefficients in [Bmpyrd][NTf2] and [Bmim][NTf2] Media abbreviation
compound
∑σx
D0,Bmpyrd (× 10-7 cm2 s-1)
D0,Bmim (× 10-7 cm2 s-1)
E1/2,1,Bmpyrd (V vs Ag/AgCl)
E1/2,1,Bmim (V vs Ag/AgCl)
BP1
4,4′- dimethoxybenzophenone
-0.54
2.3
-1.90
-1.87
BP2
4,4′-dimethylbenzophenone
-0.34
2.7
-1.81
-1.75
BP3
3-methylbenzophenone
-0.07
3.4
-1.75
-1.70
BP4
benzophenone
0.0
2.5
1.5a 1.4b 1.3a 1.1b 1.1a 0.98b 1.2a 1.0b
-1.74
-1.69
a
Where n ) 1 at high sweep rates. b When n ) 2 at low sweep rates.
Behavior of Electrogenerated Bases
J. Phys. Chem. B, Vol. 111, No. 31, 2007 9283
Figure 3. Plot of ipc1,Bmpyrd vs ν1/2 for BP2 in [Bmpyrd][NTf2].
chemical reaction consuming the electrogenerated radical anions, which is likely to be the well-known radical anion dimerization reaction. On the basis of the relative cathodic and anodic peak currents, the rate constant for this reactions is of the order 10-3 s-1. Overall, the fact that the BP1-4 radical anions are chemically stable at such low potential sweep rates suggests that these species neither become protonated, nor form strong ion-pairs, within the [Bmpyrd] medium.19,22 This lack of ionpairing is consistent with previous reports which demonstrate the absence of ion-pairing with organic cations.28 Plotting the cathodic peak currents (ipc1,Bmpyrd) for the reduction process recorded in this medium as a function of ν0.5 (as shown in Figure 3 for BP2) yields strictly linear behavior from 0.001 to 5 V s-1, which indicates that the currents are purely diffusion-controlled. This observation also confirms that radical anion protonation, or ion-pairing, within the RTIL does not occur, otherwise the protonated, or ion-paired radical anion, would coreduce with the parent BP, behavior which would be evident as sweep rate dependent cathodic currents. The diffusion coefficient (D0,Bmpyrd) for BP1-4 within this medium obtained from the slope29 of such plots (Figure 3) are given in Table 1. D0,Bmpyrd values were in the range (2.3-3.4) × 10-7 cm2 s-1, which are typical for molecular species in viscous RTILs, i.e., approximately 2 orders of magnitude smaller than in organic solvent.30 It is also noticeable in Figure 2 that the peak-to-peak separation (∆Ep) for the initial reduction process is = 0.065 V, which is slightly larger than the theoretical value of 2.303RT/ nF ()0.058 V) for a reversible redox couple29 which seems to suggest slow heterogeneous electron-transfer kinetics. However, since ∆Ep remains constant with increasing sweep rate, the slightly larger ∆Ep is likely to be the effect of some residual uncompensated solution resistance (iR); as such, it can be surmised that the heterogeneous rate constants (k0) for BP1-4 are at least 0.1 cm s-1, which is the minimum value for reversibility. These values are comparable to those reported for unsubstituted benzophenone reduction in a variety of molecular solvents at metal electrodes where values of k0 were of the order of 0.1-2.7 cm s-1 in MeCN (Au),31 DMF (Au),31 and DMSO (Hg).32 The relatively facile kinetics observed here for benzophenone reduction is consistent with those previously reported by us for benzaldehyde reduction9 and TEMPO oxidation13 in the same medium. Figure 4a,b shows CVs with extended cathodic limits recorded at 10 and 1.0 V s-1 for the reduction of BP4. It is evident that at the higher sweep rate (Figure 4a), a second reversible one-electron reduction occurs at -2.47 V vs Ag/AgCl. This reduction can be attributed to the reversible electrochemical generation of the dianion species. Since it is known that
Figure 4. Cyclic voltammograms of BP4 in [Bmpyrd][NTf2] at sweep rates of (a) 10.0 and (b) 1.0 V s-1 with extended cathodic potential limits.
protonation33 or strong (bond forming) ion-paring34 leads to irreversibility, this observation demonstrates that neither protonation of, nor strong ion-pairing with, the highly basic dianion occurs within this medium at these time-scales. It can also be observed (Figure 4a) that the second reversible reduction process is 0.62 V more negative than the parent reduction. This large potential separation (∆Epc) indicates that the thermodynamic tendency of the radical anions to comproportionate (giving the neutral parent and dianion species) is small in this medium. The equilibrium constant for comproportionation is given by eq 1,19 where Epc1 and Epc2 are the peak potentials
∆Epc ) Epc2 - Epc1 ) RT/nF ln Kcomp
(1)
for the first and second reductions, respectively. With the use of this formulation, Kcomp in the [Bmpyrd] medium is 2.1 × 10-11, which is insignificant and consistent with previous reports.33 It is interesting to compare the reversible behavior observed here with electrochemistry investigated in liquid NH3 (at 50 °C),33 where reversibility of the BP4 radical anion/dianion was observed in the presence of KI, NaI, LiNO3, and CH3 (n-Bu)3NI electrolytes. This reversibility, which is essentially the same as that observed here in [Bmpyrd][NTf2], was ascribed to the aprotic nature of the medium and the lack of strong ionpair formation between electrogenerated anions and electrolyte cations. In the NH3 medium, it was presumed that the electrogenerated anions and electrolyte cations were fully solvated such that solvent-separated ion-pairs (weak interaction) were formed. Also, the potential for the second reversible reduction in liquid NH3 was 0.53 V more negative that the parent reduction, whereas the ∆Epc separation observed in [Bmpyrd][NTf2] is 0.62 V. It is known that “moderate” ion-pairing will shift the reduction potential anodically by 59.6 mV (at 298 K) per decade change in the ion association constant.34 Therefore, the approximate -90 mV difference between ∆Epc values in [Bmpyrd][NTf2] and liquid NH3 suggests that the strength of any ionparing interaction involving the dianion in the ionic liquid medium is lower than that in liquid NH3 containing KI, NaI, LiNO3, and CH3(n-Bu)3NI electrolytes. This is probably a result of the significant charge delocalization within the [Bmpyrd] cations. Voltammetry with extended cathodic limits at the lower sweep rate of 1.0 V s-1 is shown in Figure 4b where it can be seen that the second reduction process begins to tend toward electrochemical irreversibility at the lower sweep rate, behavior
9284 J. Phys. Chem. B, Vol. 111, No. 31, 2007
O’Toole et al.
Figure 6. Plot of ipc1,Bmim vs ν1/2 for BP1 in [Bmim][NTf2].
Figure 5. Cyclic voltammograms for the first reduction of BP4 in [Bmim][NTf2] at sweep rates of (a) 0.005, (b) 0.150, and (c) 7.500 V s-1.
indicative of a slow follow-up chemical process. Since chargecompensating pyrrolidinium cations are in large excess and do not appear to form strong ion-pairs with the dianion, it is likely that the slow process is protonation of the highly basic dianion by trace H2O within the ionic liquid medium. Since it is known that a protonation reaction will result in an anodic shift in reduction potential (from the reversible potential) of 59 mV/ log[H+], the near-equivalence of E1/2 at 10.0 and 1.0 V s-1 supports the supposition that the protonation occurs from trace H2O. Electrochemistry of BP1-4 in [Bmim][NTf2]. As mentioned in the introduction, the [Bmim] cation is protic, with a reported pKa in the range 21-23,23 and BP electrochemistry is sensitive to proton availability.24 Figure 5a-c shows CVs for the reduction of BP4 in [Bmim][NTf2] at various sweep rates. It is clearly evident that at low sweep rates (Figure 5a, 0.005 V s-1), the electrochemistry involves a single chemically irreversible process. However, at increased sweep rates (to 0.150 V s-1 (Figure 5b) and 7.500 V s-1 (Figure 5c)), two closely spaced reduction processes emerge. The other significant observations
are (a) that the potential for the first reduction to generate the radical anion (as half wave potentials, E1/2,1,Bmim, Table 1) observed at higher sweep rates is only approximately 50 mV more positive in [Bmim] compared to [Bmpyrd] for BP1-4, suggesting that radical anion interaction (strong ion-pairing or protonation) with [Bmim] does not occur significantly at the faster time-scale,19,32 (b) the peak potential for the irreversible reduction (that observed at lower sweep rates, Figure 5a) is shifted positively with decreasing potential sweep rates relative to the reversible potential (observed at high sweep rates), which suggests that a slow follow-up chemical reaction occurs, and (c) the potential for the second irreversible reduction is shifted significantly positive relative to [Bmpyrd], suggesting that either a strong ion-stabilizing interaction, or protonation, of the dianion occurs within [Bmim][NTf2]. It can also be observed that the initial reduction process only tends toward partial reversibility at higher sweep rates, whereas the second reduction is strictly chemically irreversible. These observations are clearly different from those observed in [Bmpyrd][NTf2]. At first sight, the irreversibility at low sweep rates seems to indicate that the n ) 1 radical anion dimerization reaction is occurring at low sweep rates. However, plotting the cathodic current (ipc1,Bmim) as a function of ν0.5 (shown in Figure 6 for BP1) reveals two distinct linear regimes. At low sweep rates, the slope is larger than at high sweep rates. If the radical anionradical anion dimerization reaction were responsible for irreversibility at low sweep rates, the current should be a direct function of ν0.5. However, if we make the assumption that a two-electron process occurs at low sweep rates, whereas only one electron is involved at higher sweep rates, and then calculate the corresponding D0,Bmim values (given in Table 1), these are found to be effectively identical. For example, for BP1, the diffusion coefficients are (1.43 and 1.46) × 10-7 cm2 s-1 with n ) 2 and n ) 1, respectively. This indicates a transition from n ) 1 to n ) 2 upon varying the time-scale of the measurements. This observation is confirmed by chronoamperometry, which is equivalent to slow-sweep rate cyclic voltammetry in terms of electron number (n), where the transient current-time data are plotted (Figure 7) in the Cottrell form (i(t) vs t-0.5) for BP4 reduction. If it is assumed that the D0,Bmim (n ) 1) values obtained from the high sweep rate CVs are accurate descriptions of mass transport, application of the Cottrell equation reveals that n ) 2.0 ( 0.02 for BP1-4 in the [Bmim] medium. It is, therefore, clear that a slow chemical process occurs after the initial electron-transfer that results in the transition from a oneelectron process to a two-electron process upon increasing the time-scale of the experiment. The origin of this behavior may be due to either (a) coreduction of the parent BP along with the
Behavior of Electrogenerated Bases
Figure 7. Cottrell plot for the reduction of BP4 in [Bmim][NTf2].
SCHEME 1: The ECE and DISP Mechanisms for the Hypothetical Two-Electron Reductions of Species A-D
radical anion to give a fully protonated/stabilized ion-paired (from/with [Bmim]) dianion (i.e., merging of the two reduction processes) or (b) the involvement of the radical anion in a slow protonation or stabilizing ion-pairing process with the [Bmim] cation19,22 ultimately leading to an electrochemical, chemical, electrochemical-type (ECE) two-electron process. In order to investigate the reason for the n ) 1 to n ) 2 transition, it is useful to review similar observations in the literature. It has been shown24,35 that the reduction of unsubstituted benzophenone in aqueous media at near-neutral pH (pH 6-8) exhibits an irreversible two-electron reduction similar to that observed here in [Bmim][NTf2]. This behavior was attributed to an ECE-type mechanism involving protonation of the radical anion (pKa ≈ 9.236) leading to the neutral radical which is coreduced with the parent molecule ultimately giving rise to an n ) 2 process, overall. Alternatively, Save´ant et al.36 and Fujihira et al.37 reported a disproportionation-type (DISP) reaction after the initial BP reduction in the presence of proton donors, which also leads to a two-electron process involving the protonation of the radical anion. Also, evidence in the literature28 suggests that BP radical anions do not ion-pair with organic cations (in molecular solvents at least). This information indicates that BP1-4 undergo either an ECE-type or a DISPtype process, via a protonated radical anion, in [Bmim] RTIL. The ECE and DISP mechanisms are shown in Scheme 1 for the hypothetical two-electron reductions of species A-D. Now, DISP-type mechanisms can be classified according to the relative rates of the protonation (kp) and disproportionation (kd) reactions.36 If the first-order radical anion protonation reaction (kp) is sufficiently fast for equilibrium to be maintained at any potential, then the second-order disproportionation process (kd) is the rate-determining step, in which case the reaction is classified as the DISP2 case. Alternatively, if kp is ratedetermining, this is classified as the DISP1 case. These cases can be distinguished36 by plotting the reduction potential vs log υ where a slope of -19.4 mV per decade corresponds to the DISP2 case, whereas a slope -29.1 mV indicates either an ECE or a DISP1 mechanism. Unfortunately, as pointed out by Save´ant et al.,36 ECE and DISP1 processes are thermodynamically equivalent and can only be distinguished under certain circumstances.38 Notwithstanding, we can easily distinguish between
J. Phys. Chem. B, Vol. 111, No. 31, 2007 9285
Figure 8. Save´ant diagnostic plot of Epc1,Bmim vs log ν for BP3.
either the ECE/DISP1 cases and the DISP 2 case using the aforementioned diagnostics. Figure 8 shows a plot of Epc1,Bmim vs log υ for BP3 where the slope, over the sweep rate range corresponding to irreversibility, is -27 mV decade-1. For BP1, BP2, and BP4, the slopes are -29, -26, and -32 mV decade-1, respectively. From this, it is apparent that either an ECE or DISP1 mechanism occurs in [Bmim]. Irrespective of which mechanism, this implies that the radical anion slowly abstracts protons from the [Bmim] medium and that this protonation reaction is the rate-determining step. Since the ionic liquid medium is effectively anhydrous, this observation implies that the protic [Bmim]23 is the source of protons. This is consistent with other studies which have shown that deprotonation of [Bmim] can occur in the presence of relatively weak bases (pKa 8-9),39 while Muldoon et al.40 have shown that the photoexcited state of BP abstracts protons from [Bmim] to form the neutral radical, BPH•. It is worth noting that deprotonation of [Bmim] results in the formation of the carbene species, which is highly unstable. It is also worth noting that, although the relative pKa’s for the radical anion/ketyl radical (≈9.2 for benzophenone radical anion/neutral radical in DMF36) and [Bmim]+/[Bmim]- (≈212323) dictate that only a very low equilibrium concentration of the neutral radical can exist, which is too low to result in the accelerated n ) 2 cathodic current, and it is the rate of radical anion protonation that is sufficiently fast to ensure the twoelectron behavior observed at low sweep rates.41 From observation, we know that BP1, BP2, BP3, and BP4 reductions begin to become reversible at approximately 0.5, 0.4, 0.2, and 0.1 V s-1, respectively. With this information, the rate constants for the protonation reaction (kp) have been evaluated for each BP by digital simulation of the corresponding cyclic voltammograms. In these simulations, an ECE mechanism was used where kp was varied while keeping the sweep rate constant (i.e., 0.5, 0.4, 0.2, and 0.1 V s-1) until the reoxidation process begins to appear. The corresponding approximate rate constants of the protonation reaction of BP1, BP2, BP3, and BP4 ketyl radical anions are 0.12, 0.10, 0.045, and 0.021 s-1, respectively. This indicates that the rate of the proton-transfer reaction increases with decreasing Hammett number (Table 1), which is a simple reflection of the increased electron density upon, and therefore the basicity of, the electrogenerated radical anion. Now, the Hammett free-energy relation42 describing the substitutents’ effects on the rate of proton transfer, relative to the unsubstituted species, is given by
log kp,x ) log kp,0 - (F
∑σx)
(2)
9286 J. Phys. Chem. B, Vol. 111, No. 31, 2007
O’Toole et al.
Figure 9. Plots of log(kf/k0) vs ∑σx for BP1-4 reduction in [Bmim][NTf2] (solid line) and theoretical curve (dashed line).
where kp,x is the protonation rate constant for the substituted BP1-3, kp,0 is the protonation rate constant for unsubstituted (BP4) radical anion (0.021 s-1), F is the Hammett solventdependent reaction constant (it will be shown below that F ) 0.307 in [Bmim][NTf2]), and ∑σx is the relevant Hammett substitutent constant. A plot of log(kp,x/kp,0) vs -∑σx is shown in Figure 9a. The expected strict linearity of the data is clearly not evident (and is not expected to be, given the approximate nature of the rate constant data); however, forcing linear regression through the origin yields a slope of 0.30, which compares favorably with the theoretical slope of 0.307, as shown in Figure 9b. This suggests that the deprotonation of [Bmim] by the radical anion follows a simple free-energy relation. From this, the pKa’s of the radical anions (BP1-3), relative to BP4 in [Bmpyrd][NTf2], can be approximated, which are 9.74, 9.54, and 9.27, respectively. The observation made here that electrogenerated organic radical anions can deprotonate [Bmim] introduces the question as to why the highly reactive superoxide species is stable14-16 in protic imidazolium-based ionic liquids. This can be explained easily by considering the relative pKa’s of the superoxide species and the electrogenerated radical anions studied here. The pKa for the reaction HO2• h O2•- + H+ is 4.7,43 which is 4.5 orders of magnitude smaller than that for BP4. Since the rate of [Bmim] deprotonation is governed by the basicity of the electrogenerated radical anion, the pKa of the superoxide species is sufficiently small to ensure that the rate of [Bmim] deprotonation is unobservable at the lowest time-scales (9 mV s-1) reported previously.14 RTIL Polarity. It is well-known that the half-wave potential for the first redox process (radical anion generation) of substituted benzophenones (E1/2,x) is a strict function of their Hammett substituent constants44,45 and adheres to the following free-energy relationship (eq 3):
∑σx
E1/2,x - E1/2,0 ) ∆E1/2 ) F
(3)
where E1/2,0 is the half-wave potential for the first reduction of the unsubstituted benzophenone (BP4). As mentioned above, the slope, F, is the solvent-dependent reaction parameter, the value of which varies depending on the polarity of the solvent.25 The half-wave potentials for BP1-4 (benzophenone), and three other substituted benzophenones (not discussed here), are plotted as a function of ∑σx in Figure 10a,b for [Bmpyrd][NTf2] and [Bmim][NTf2], respectively. Obviously, the plots are linear which indicates that the nature of the substituent has no differential effect on the BPs’ interactions with the medium. The slopes for these plots are 0.320 and 0.307 for [Bmpyrd]
Figure 10. Hammett plots of -Epc1 vs ∑σx for reduction in (a) [Bmpyrd][NTf2] and (b) [Bmim][NTf2]. Points corresponding to ∑σx values of 0.23, 0.46, and 1.0 are for the reduction of 4-bromobenzophenone, 4,4′-dichlorobenzophenone, and 4-cyanobenzophenone, respectively.
and [Bmim] ionic liquids, respectively. The near equivalence of these values indicates that the polarity of both media are very similar, as observed elsewhere.46 Also, these values are similar to polar solvents such as acetonitrile47 where F ) 0.33, which is in keeping with other reports of RTIL polarity.46,48,49 Conclusions The electrochemical reduction of substituted benzophenones in aprotic [Bmpyrd][NTf2] occurs via two consecutive oneelectron processes leading initially to the radical anion and subsequently to the dianion species. The radical anion in this medium is highly stable at all measurement time-scales while the voltammetry of the dianion species is reversible at moderately low sweep rates of g10.0 V s-1. These observations, along with the observation that dianion generation occurs 0.62 V more cathodic than the parent reduction, indicate that radical anion and dianion ion pairing with the [Bmpyrd] cation is weak, or nonexistent. Also, the reduction potential for the generation of the dianion occurs at potentials not accessible with conventional aprotic solvents. Collectively, these observations indicate that the [Bmpyrd][NTf2] RTIL is a superb medium for fundamental investigation of electrochemical (and coupled chemical) processes involving highly reactive electrogenerated intermediates. In contrast, the reductive electrochemistry of substituted benzophenones in the protic [Bmim][NTf2] depends on the timescale of the measurements. At sweep rates on the V s-1 timescale, the initial reduction (n ) 1, radical anion generation) is partially reversible whereas the second reduction process (n ) 1, dianion generation) is irreversible and shifted anodically, presumably due to strong ion-pairing or protonation of the electrogenerated dianion. However, at low sweep rates (e0.1 V s-1), reduction occurs via a two-electron process with an anodic peak potential shift characteristic of an ECE- or DISP1type mechanism involving the rate-determining protonation of the radical anion via a slow proton-transfer reaction from the [Bmim] cation. The rate of [Bmim] deprotonation appears to be a function of the basicity of the electrogenerated radical anions and conforms to the Hammett linear free-energy relation. This observation also explains the stability of the superoxide radical anion in [Bmim]-based ionic liquids. Collectively, these observations indicate that electrochemistry involving the formation of electrogenerated bases in imidazolium ionic liquids, at long time-scales, is influenced by proton-transfer events and that the unstable carbene formation will occur.
Behavior of Electrogenerated Bases The half-wave potentials for the initial one-electron reduction of substituted benzophenones in both [Bmpyrd][NTf2] and [Bmim][NTf2] are a direct function of the corresponding Hammett substituent constants which indicates that all solutes interact equivalently with the respective media. Also, the near identical slopes (reaction parameters) indicated that both liquids exhibit similar polarity, which is essentially equivalent to polar molecular solvents. Acknowledgment. S.O’T. thanks Queen’s University for an APG studentship. References and Notes (1) Welton, T. In Ionic Liquids in Synthesis; Wasserscheid, P., Welton, T., Eds.; Wiley, VCH: Weinheim, Germany, 2003. (2) Welton, T. Chem. ReV. 1999, 99, 2071. (3) Enthaler, S.; Jackstell, R.; Hagemann, B.; Junge, K.; Erre, G.; Beller, M. J. Organomet. Chem. 2006, 691, 4652. (4) Mo, J.; Ruan, J.; Xu, L. J.; Hyder, Z.; Saidi, O.; Liu, S. F.; Pei, W.; Xiao, J. L. J. Mol. Catal. A: Chem. 2007, 261, 275. (5) Miao, S. D.; Liu, Z. M.; Han, B. X.; Huang, J.; Sun, Z. Y.; Zhang, J. L.; Jiang, T. Angew. Chem., Int. Ed. 2006, 46, 266. (6) Migowski, P.; Dupont, J. Chem.sEur. J. 2007, 13, 32. (7) Doherty, A. P.; Brooks, C. A. In Ionic Liquids as Green SolVents: Progress and Prospects; Rogers, R. D., Seddon, K. R., Eds.; ACS Symposium Series 856; American Chemical Society: Washington, DC, 2003; pp 410-420. (8) Eisele, S.; Schwarz, M.; Speiser, B.; Tittel, C. Electrochim. Acta 2006, 51, 5304. (9) Doherty, A. P.; Brooks, C. A. Electrochim. Acta 2004, 49, 3821. (10) Barhdadi, R.; Courtinard, C.; Nedelec, J. Y.; Troupel, M. Chem. Commun. 2003, 1434. (11) Earle, M. J.; McCormac, P. B.; Seddon, K. R. Green Chem. 1999, 1, 23. (12) Rencurosi, A.; Lay, L.; Russo, G.; Caneva, E.; Poletti, L. J. Org. Chem. 2005, 70, 7765. (13) Barhdadi, R.; Comminges, C.; Doherty, A. P.; Nedelec, J.-Y.; O’Toole, S.; Troupel, M. J. Appl. Electrochem. 2007, 37, 723-728. (14) AlNashef, I. M.; Leonard, M. L.; Kittle, M. C.; Matthews, M. A.; Weidner, J. W. Electrochem. Solid-State Lett. 2001, 4, D16. (15) AlNashef, I. M.; Leonard, M. L.; Matthews, M. A.; Weidner, J. W. Ind. Eng. Chem. Res. 2002, 41, 4475. (16) Martiz, B.; Keyrouz, R.; Gmouh, S.; Vaultier, M.; Jouikov, V. Chem. Commun. 2004, 674. (17) Lagrost, C.; Carrie, D.; Vaultier, M.; Hapiot, P. J. Phys. Chem. A 2003, 107, 745. (18) Brooks, C. A.; Doherty, A. P. J. Phys. Chem. B 2005, 109, 6276.
J. Phys. Chem. B, Vol. 111, No. 31, 2007 9287 (19) Fry, A. J. J. Electroanal. Chem. 2003, 546, 35. (20) Lagrost, C.; Gmouh, S.; Vaultier, M.; Hapiot, P. J. Phys. Chem. A 2004, 108, 6175. (21) Lagrost, C.; Hapiot, P.; Vaultier, M. Green Chem. 2005, 7, 468. (22) Lagrost, C.; Preda, L.; Volanschi, E.; Hapiot, P. J. Electroanal. Chem. 2005, 585, 1. (23) Amyes, T. L.; Diver, S. T.; Richard, J. P.; Rivas, F. M.; Toth, K. J. Am. Chem. Soc. 2004, 126, 4366. (24) Zuman, P.; Barnes, D.; Ryvolova-Kejharova, A. Disc. Farad. Soc. 1968, 45, 202. (25) Zuman, P. In Substituent Effects in Organic Polarography; Plenum Press, New York, 1967, pp 23-41. (26) MacFarlane, D. R.; Meakin, P.; Sun, J.; Amini, N.; Forsyth, J. J. Phys. Chem. B 1999, 103, 4164. (27) Isse, A. A.; Gennaro, A. Collect. Czech. Chem. Commun. 2003, 68, 1379. (28) Kalinowski, M. K. Chem. Phys. Lett. 1970, 7, 55. (29) Brett, C. A. M., Brett, A. M. O. Electrochemistry, Principles, Methods and Applications, 2nd Ed., Oxford Science Publications, 1993 pp 174-198. (30) Evans, R. G.; Klymenko, O. V.; Price, P. D.; Davies, S. C.; Hardacre, C.; Compton, R. G. ChemPhysChem 2005, 6, 526. (31) Walczyk, K. R.; Popkirov, G. S.; Schindler, R. N. Ber. Bunsenges. Phys. Chem. 1995, 99, 1028. (32) Fawcett, W. R.; Fedurco, M. J. Phys. Chem. 1993, 97, 7075. (33) Demortier, A.; Bard, A. J. J. Am. Chem. Soc. 1973, 95, 3495. (34) Save´ant, J. M. J. Phys. Chem. B 2001, 105, 8995. (35) Saraswathi, R.; Narayan, R. Proc. Indian Acad. Sci. (Chem. Sci.) 1986, 97, 403. (36) Nadjo, L.; Save´ant, J. M. J. Electroanal. Chem. 1971, 33, 419. (37) Fujihira, M.; Suzuki, H.; Hayano, S. J. Electroanal. Chem. 1971, 33, 393. (38) Compton, R. G.; Mason, D.; Unwin, P. R. J. Chem. Soc., Faraday Trans. 1988, 84, 2057. (39) Aggarwal, V. K.; Emme, I.; Mereu, A. Chem. Comm. 2002, 1612. (40) Muldoon, M. J.; McLean, A. J.; Gordon, C. M.; Dunkin, I. R. Chem. Comm. 2001, 2364. (41) Andrieux, C. P.; Merz, A.; Save´ant, J. M.; Tomahogh, R. J. Am. Chem. Soc. 1984, 106, 1957. (42) Hammett, L. P. Chem. ReV. 1935, 17, 125. (43) Singh, A. Can. J. Physiol. Pharmacol. 1982 60 1130. (44) Cauquis, G. In Organic Electrochemistry - An Introduction and a Guide, Baizer, M. M., Ed., Marcel Dekker Inc: New York, 1973, p. 79. (45) Zuman, P. In Substituent Effects in Organic Polarography, Plenum Press, New York, 1967, pp 23-41. (46) Mandal, P. K.; Samanta, A. J. Phys. Chem. 2005, 109, 15172. (47) Jaworski, J. S.; Malik, M.; Kalinowski, M. K. J. Phys. Org. Chem. 1992, 5, 590. (48) Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W. J. Am. Chem. Soc. 2002, 124, 14247. (49) Karmakar, R., Samanta, A. J. Phys. Chem. 2002, 106, 6670.
